Reclamation of steeply sloping coal spoil banks under Mediterranean semi-arid climate.
Mining activities can cause severe environmental impact and degradation if reclamation measures are not taken. Deep coal mining activities in the Ebro Valley (NE Spain) are important elements in the economy of the region. Mine spoils are deposited on the surface, changing the natural topography through creation of hills, slopes, and valley infill.
Usually on the slopes of the new topography, erosion is a serious problem, and the Mediterranean climatic conditions are especially favourable to the development of erosion processes. Erosion of mine spoils may result in a decreased land use capability and sustainability, landform instability, and degradation of downstream water quality (Loch 2000a; Evans 2000). The control of erosion is therefore a principal concern because of its impact on landscape stability.
Due to the fact that erosion rates are expected to increase with slope gradient, mine reclamation includes recontouring to low slopes. These operations represent a major cost in rehabilitation of mined lands (Carroll et al. 2000; Sheridan et al. 2000a); however, steep topography and limited space may require steep slopes or slopes with the angle of repose of the materials used in their construction.
In such cases, revegetation of slope banks is a high priority because it has proven to be effective in reducing runoff volume and substrate loss (Wilcox and Wood 1989; Carroll et al. 2000) as well as runoff pollutants (Sheridan et al. 2000b).
Several factors can affect the establishment of vegetation on slopes by influencing seed migration and fixation and seedling germination. Physical forces such as wind, and slope characteristics like gradient or length, may cause seed removal. When these factors interact with seed and soil surface characteristics, such as roughness or surface strength, they may influence the distance and direction of seed dispersal (Johnson and Fryer 1992). Observations of naturally vegetated mine slopes have shown higher plant cover and species richness at the base of slopes than the upper and mid slopes, indicating seed removal and migration.
Other studies have found that seed losses are very low on slopes (Garcia-Fayos et al. 1995). In such cases, the absence of vegetation may be the result of seedling mortality or non-germination, due to an unsuitable moisture regime within the soil (Cerda and Garcia-Fayos 1997) or the absence of microsite conditions that favour seed germination (Leavitt et al. 2000).
Different factors affect runoff characteristics. Erosion rates are generally assumed to increase with greater slope length or gradient; however, some experiments have shown that some soils exhibit no increase in erosion rates with increasing slope length (Loch 1996), which, in some cases, cannot be considered a major factor affecting erosion (Evans 2000) or slope gradient (Schroeder 1987).
This study examines the feasibility of revegetation of mine dumps in the Mequinenza coalfield, many of which are barren and eroded. An adequate substrate is essential for the development of a good vegetation, but lack of topsoil is a problem at many sites globally (because topsoil is not collected and salvaged, or it has a poor quality) and this area is no exception.
Fortunately, farm by-products rich in organic matter are easily available in the zone. These local resources can be used to improve mine spoil properties, and amended spoils can be used as topsoil substitutes. Applying sewage sludge (Sopper 1992), municipal solid wastes (Felton 1995), or farm slurries (Hornick 1988; Leiros et al. 1996; Ye et al. 1999) to mine spoils has proven to be effective in enhancing revegetation. Among available byproducts in this area there are pig slurry and barley straw. The application of pig slurry, however, may have adverse side effects, mainly if nitrates (N[O.sub.3.sup.-]-N) run off and pollute the water (Smith et al. 1990).
The purpose of this study was to find a method of reclamation and revegetation on high slope spoil banks of coal mines in semi-arid Mediterranean climate conditions by using amended mine spoils (with pig slurry and straw) as a substrate. A second aim was to determine the main variables affecting the establishment of vegetation. Factors considered included slope characteristics, substrate characteristics, and plant species.
Materials and methods
The Carbonifera del Ebro coalmine company is located in the Mequinenza coalfied, between south-west Lleida province and south-east Zaragoza, where the mine occupies 498 ha. Lignite is extracted by deep mining (chamber galleries), producing 500 000-1 000 000 t of lignite per year with an efficiency of only 12%. Overburden is deposited either by filling valley bottoms, reaching 25 m in thickness, or on terraced slopes, which are up to 100 m in height with several terrace levels.
The climate is semi-arid Mediterranean with an average annual temperature of 14[degrees]C, average annual rainfall of 395 mm, and average Thornthwaite annual evapotranspiration of 950 mm. There is a substantial water deficit, mainly during July and August. The area is also affected by a persistent and forceful wind coming from the west, whose effects are reinforced by the aspect of the dump at the study site (mainly west aspect). Geological materials of the area are Tertiary lacustrine limes and gypsum alternating with continental marls, mostly saline.
The dump where the experimental plots were installed was re-contoured from the initial angle of repose of the materials in 1995. The steep topography and limited space in the area called for high gradient slope batters (>75% in some of them). The base of the dump was limited by the Mequinenza Reservoir on the Ebro River, which supplies drinking and irrigation water to a large area. Three years after the recontouring of the spoil, there was a total lack of vegetation and the development of active erosion features as rills, gullies and mass movements (Fig. 1).
[FIGURE 1 OMITTED]
Sixteen experimental plots sized from 7.0 by 13.0 m to 7.0 by 17.5 m (the long axis of each plot was up-and-down the slope) were established on a slope spoil bank (re-contoured by the mining company) in October 1998, Heavy earth moving equipment was used to construct the plots. The plots were organised in 4 sets (Sets I-IV) according to slope aspect, slope gradient and plot length. Each set had 4 plots, with 2 replications (randomly distributed within each set) of two different substrate types, designated type 1 : 1 and type 1 : 2. A 1-m minimum buffer between plots completed each set (Fig. 2).
[FIGURE 2 OMITTED]
Two types of overburden, namely fine and coarse spoil, mixed with pig slurry and straw, were used as raw materials for the two types of substrate (Table 1). Fine spoil was a silty material from lagoon sedimentation resulting from the process of coal washing, without coarse fragments. Coarse spoil was the sieving remnant, containing 89% by weight coarse fragments (93% of these 2-6 cm), mainly limestone mixed with gypsum and lignite (this is the same spoil employed by the mine company to construct the banks).
In August 1998, fine spoil, stored on a dried lagoon, was uncompacted leaving a 10-cm layer. Straw at 9 t/ha was then applied. Pig slurry was spread weekly, with 3 additions totalling 370 [m.sup.3]/ha. In October 1998, another application of straw and pig slurry was made at 12 t/ha of straw and 495 [m.sup.3]/ha of pig slurry. After that, crossed-ploughing with a chisel down to 10 cm depth was performed, concluding the preparation of the fine amended spoil.
The 1 : 1 substrate type was made by combining the same amount by volume of fine amended spoil as coarse spoil. The 1 : 2 substrate type had double the volume of coarse spoil as fine amended spoil. Each plot was made using 24 [m.sup.3] of substrate, which is why the plots reached depths of 20 or 40 cm only depending on their length 17.5 or 13.0, m respectively.
Four seed mixtures were hand-broadcasted in [1-m.sup.2] area microplots, where the surface material was slightly ripped prior to sowing. Sowing combinations were:
(i) R-mixture: mixture of indigenous saline and semi-arid adapted bushes and grasses, 300 seeds/[m.sup.2]; equal proportions of Atriplex halimus L., Bassia scoparia (L.) Voss, Medicago sativa L., Oryzopsis milliacea L., Salsola kali L., and Salsola vermiculata L.
(ii) A-mixture: mixture of indigenous semi-arid adapted bushes, 170 seeds/[m.sup.2]; proportions--Globularia maritima L. 30%, Pistacia lentiscus L. 5%, Retama sphaerocarpa L. 5%, Rosmarinus officinalis L. 30%, and Thymus vulgaris L. 30%.
(iii) T-mixture: species and proportions the same as R-mixture, 600 seeds/[m.sup.2].
(iv) F-mixture: commercial mixture of semi-arid adapted hydrosowing grasses, 30 g/[m.sup.2]; proportions--Agropyron repens (L.) Beauv. 15%, Dactylis glomerata L. 25%, Festuca arundinacea Schreber 20%, Lolium rigidum Gaudin 25%, Melilotus officinalis (L.) Pallas 5%, and Onobrychis viciiefolia Scop 10%.
In each plot the R- and A-mixtures, and a non-sown microplot to monitor spontaneous colonisation (B-microplot), were placed, in December 1998, 1.5 m from the top and bottom boundaries of each plot, and the mixtures were randomly distributed within the upper and lower parts of each plot. T- and F-mixtures were sown in March 1999, on the mid-lower parts of each plot, separated 3 m from the bottom limit of the plot and randomly distributed (Fig. 3).
[FIGURE 3 OMITTED]
Seeds of Atriplex halimus, Bassia scoparia L., Globularia maritima L., Oryzopsis milliacea L., Pistacia lentiscus L., Retama sphaerocarpa L., Salsola kali L., and Salsola vermiculata L. were collected in November 1998; the rest of the seeds were purchased. Laboratory germination experiments were carried out to check collected (not purchased) seeds viability using a Jacobsen germinator (Besnier 1989).
Surface strength was measured in February 1999 using the spring penetrometer developed by Nacci and Pla (1992). Nine measurements were taken in the 3 upper and 3 lower microplots of each experimental plot; 54 measures were taken in each plot, in dry surface conditions (average substrate moisture was <2% during the determination period).
Superficial moisture (top 2 cm) was measured 48 h after a rainfall event during the emergence period in the upper (sampled at 0.5 m from the top of the plot) and lower parts (sampled at 0.5 m from the bottom of the plot) of each plot. Bulk samples were taken, and kept hermetically sealed to avoid moisture loses.
Nitrate concentration and substrate salinity of the top 20 cm were monitored from February 1999 to March 2000. Sampling strategy was the same as used in monitoring substrate moisture. Bulk samples were taken, kept hermetically sealed to avoid moisture losses, and placed in a portable cool box to avoid nitrogen transformation reactions during transportation. On arrival at the laboratory, samples were kept refrigerated prior to analysis.
Samples were weighed then sieved through a 2-mm sieve. A fraction of fine earth was used to determine the N[O.sub.3.sup.-]-N concentration (mg/L extract basis) using the Merckoquant test strips and a Nitracheck reflectometer. Nitrates were extracted with water, and samples were compared with standard solutions analysed at the same temperature to minimise errors (Wetselaar et al. 1998). Electrical conductivity (EC) was determined using a 1 : 1 soil to water extract (dS/m at 25[degrees]C) and was assessed using a conductivity meter (Porta et al. 1986). To obtain the moisture content, the resting sample was dried to a constant weight, at temperature <40[degrees]C to avoid water losses owing to gypsum (Vieillefon 1979). Moisture and Nitracheck[R] values were employed to calculate the N[O.sub.3.sup.-]-N concentration in mg/kg on a dry-fine earth basis.
Daily precipitation was measured using a pluviometer model Rain-o-Matic[R]. Sediment yield, N[O.sub.3.sup.-]-N concentration, and EC were measured in runoff water samples from Gerlach sediment boxes 0.5 by 1.0 m wide, installed at the bottom of each plot. Samples were collected from January to June 1999 following major storms. In the period of study, data were recovered after 7 rainfall events. Water samples were removed from the collecting bottles (a fraction of maximum 2 L was taken from each bottle after vigorous agitation to homogenise the sample), kept in hermetic recipients, and treated as for the substrate samples prior to analysis.
Sediment concentration (g/L) was measured by drying and weighing, N[O.sub.3.sup.-]-N concentration (mg/L) using the Nitracheck[R], and EC (dS/m at 25[degrees]C) using a conductivity meter.
The number of emerged plants was counted in the upper and lower R, A and B microplots. Germination was counted weekly in the whole area of each microplot until it finished. After emergence, total percentage cover plant and percentage cover of individual species were monitored approximately once per month in all microplots (R, A, B, T, F). The visual rating system of Folk's diagrams (Folk 1951) was used to estimate percentages.
Statistical analyses were performed using the SAS v. 6.12 software (SAS Institute 1989). The analysis of variance (ANOVA) procedure, with Duncan's multiple range test for means separation and least squares means for interaction analysis, were employed.
Results and discussion
The substrate surface strength ranged from 1.0 to 21.6 kg/[cm.sup.2]. In Sets III and IV surface strength was higher than in Sets I and II (Fig. 4). Substrate crusting was therefore a problem in the plots, as occurs in mine spoils (Schroeder 1987; Carroll et al. 2000; Loch 2000b); however, these results are probably partly affected by the presence of stones, which increase surface strength measures made using penetrometers (Campbell and O'Sullivan 1991). Texture of the substrate, in which silt was the dominant fraction (approximately 33% in substrate type 1 : 1 and 18% in substrate type 1 : 2), seemed to be responsible for substrate crusting. According to Sheridan et al. (2000b), overburdens with 20-30% silt tended to form strong raindrop impact crusts under rainfall.
[FIGURE 4 OMITTED]
Within sets of plots (Table 2) the influence of the type of substrate varied depending on the set, without any clear tendency. Position (up or down) was important, and had a clearer effect, as the lower parts were, in general, more crusted than the upper ones. This could be due to the traffic of machinery during the plot configuration.
The superficial moisture content after rain during the emergence period ranged from 2.1 to 9.6%. Moisture content was higher in Set I than Sets II and IV (Fig. 5). The slope gradient, which conditioned runoff, and the slope aspect determined these values; southerly aspect or high slope gradient were linked to lower moisture contents. Within sets of plots there were no differences due to the type of substrate, but position had effect--Sets III and IV had greater % soil moisture in the lower parts than the upper ones (Table 3). The effect of the slope gradient was again apparent; there were no differences due to position in plots slopes near 30% (Sets I and II), but when the slope was higher the lower positions had more soil moisture than the upper ones, due to water movement from upper to lower power positions.
[FIGURE 5 OMITTED]
The N[O.sub.3.sup.-]-N substrate concentration ranged from 10 to 453 mg/kg (dry fine earth basis). Variation in time in relation to daily rainfall is shown in Fig. 6. The N[O.sub.3.sup.-]-N concentration was not different among the sets of plots in February 1999 (day of year 32). It increased quickly from this date on, probably related to an increase in the temperature, which increased the microbial nitrification activity.
[FIGURE 6 OMITTED]
The maximum N[O.sub.3.sup.-]-N concentration was reached in August (day of year 223) in Sets I and II, and in June (day of year 165) in Sets III and IV. These results are in agreement with the results of Coyne et al. (1998) in a coal mine soil reclaimed with organic wastes, where the highest mineralisation rates were found the summer month of July related with the temperature distribution along the year.
Values decreased then until March 2000, at which point there were no differences among the sets of plots. This decrease was may be due to plant absorption, immobilisation, and certain losses (mainly runoff). Absorption can be one of the main causes of N[O.sub.3.sup.-]-N decrease; amounts of nitrogen recovered by crops after applications of pig slurry may reach values of 40% (Bernal et al. 1993) or 35% (Cameron et al. 1995). In March 2000 (day of year 433), the N[O.sub.3.sup.-]-N concentrations did not differ among sets of plots; therefore, it seemed that after a year the production and different N[O.sub.3.sup.-]-N extractions and losses tended to balance its concentration in the plots.
The effect of substrate type and position within sets of plots on N[O.sub.3.sup.-]-N concentration is displayed in Table 4. Type of substrate greatly affected N[O.sub.3.sup.-]-N substrate concentration, being higher in the 1 : 1 substrate type than in the 1 : 2 substrate type for all the sets. This was because the proportion of amended fine spoil, richer in pig slurry and therefore richer in nitrogen, was higher in the 1 : 1 substrate type.
N[O.sub.3.sup.-]-N concentration in Sets I and IV was lower in upper positions than in lower positions, and the opposite happened in Sets II and Ill. The effect of position on N[O.sub.3.sup.-]-N substrate concentration seemed to be related to the slope length, a factor that also affected substrate thickness.
Substrate EC ranged from 2 to 20 dS/m. Mean variation in time, together with daily rainfall, is displayed in Fig. 7. Differences among sets of plots varied greatly. EC tended to increase with time, reaching maximum values in August (day of year 223), Set II having the highest values. These trends probably resulted from capillary ascension of salts and concentration of the soil solution. This concentration could also explain why the maximum EC value was reached in August in Set II, which had a southerly aspect and therefore higher evapotranspiration rate. From this point on EC decreased until March 2000 (day of year 433), when there were no differences among the sets of plots.
[FIGURE 7 OMITTED]
The effect of substrate type and position on EC within sets of plots (Table 5) shows that EC in substrate type 1 : 1 was higher than in substrate type 1 : 2 for the Sets I, II, and III, due to the higher slurry dose, but not Set IV. Probably the higher slope of Set IV induced lower infiltration and higher runoff, which overruled the effects of slurry dose.
Regarding positions, there were no differences between plots in Sets I and IV, but EC in Sets II and III was higher in lower positions than in upper positions. In Set II there was an interaction in which there were no differences in the 1 : 1 substrate type plots between upper and lower positions, and in lower positions there were no differences between the substrate types. These trends were difficult to explain and also reversed for N[O.sub.3.sup.-]-N concentrations. Leavitt et al. (2000) found that downslope movement of nutrients (nitrogen, phosphorous, potassium) was not detected in 30-m-long and 80% slope plots in a mine reclamation experiment using coversoil and/or fertiliser. Moreover, other processes can also affect the N[O.sub.3.sup.-]-N concentration in the substrate, such as nitrogen absorption by plants or microorganisms, denitrification, or nitrification, in such a way that is does not behave like an inorganic solute. Consequently, further research is needed to determine solute dynamics within the plots.
Erosion rills were visible on all plots; this was caused by the steep slope in all the sets of [is greater than or equal to] 30%. In Sets III and IV small mass movements showed a loss of stability when slope increased. Sheridan et al. (2000b) found rilling on 20% slope plots and little or no rilling on 5% or 10% slope plots, and Leavitt et al. (2000) found mass movements in plots established in 80% gradient slope mine dumps. Development of rills affected the sediment box measures because they increased the variability diverting or concentrating runoff depending on the intensity of the rain. Because of that a simpler analysis was done considering only the averaged values of runoff parameters during the monitoring period (Fig. 8).
[FIGURE 8 OMITTED]
Runoff sediment yield ranged from 0.8 to 136.0 g/L; the eroded material appeared to come mainly from the fine spoil, since texture and colour were similar. Sediment concentration values were not atypical for bare mine reclamation plots; e.g. Loch (2000b) describes sediment concentrations in runoff ranging from temporary peaks as high as 402 g/L for bare plots. Averaged substrate loss values in this experiment were not acceptable--it may suppose erosion rates near 100 t/ha.year.
The highest sediment concentration was measured in Set IV, which was significantly different from Set I which had the smallest substrate losses. The effect of the slope on substrate loss showed that increasing gradients meant increasing runoff sediment yield, although differences were only found between Sets I (30% slope) and IV (56% slope).
Surface conditions can greatly affect erosion and override the influence of factors more commonly expected to control erosion rates (Evans and Loch 1996). Crusting, which was observed in all plots, has shown contrasting effects on runoff. According to Sheridan et al. (2000b), substrates which tended to form strong raindrop impact crusts had low erodibilities. Carroll et al. (2000), on the other hand, describe low infiltration and large runoff and erosion rates when crusting; and Evans and Loch (1996) conclude that increased compaction and lower infiltration are the controlling factors causing higher erosion rates despite the site's lower slope (2.8% v. 20.7%). In this experiment substrate crusting should decreased the infiltration, increasing runoff rates and therefore erosion. Substrate crusting could also minimise the differences of sediment loss among the sets of plots, as was found by Schroeder (1987), who reported little difference in erosion rates from different slope gradients when spoils were crusted.
Nitrate concentration in runoff ranged from <1 to 468 mg/kg, with the highest concentration in Set I, higher than in Sets II, III, and IV, which were not significantly different from each other. This may be related to the substrate N[O.sub.3.sup.-]-N concentration measured in the Sets. It was observed that the runoff volume recovered was directly proportional to the slope gradient, so a concentration effect in the samples could also cause these results. Electrical conductivity in runoff ranged from 0.7 to 13.3 dS/m and differences among sets were not significant.
The effect of substrate type on runoff sediment yield, N[O.sub.3.sup.-]-N concentration, and EC within sets of plots is displayed in Table 6. In general, substrate type tended to influence sediment yield (except in Set II, which was south-facing); however, the effect was significant only in plots of Set IV, where the 1:1 substrate type plots were more susceptible to loss of material than the 1:2 substrate type plots. Over all sets, the 1:1 substrate type averaged 40.8 g/L, whereas the 1:2 type averaged 25.2 g/L. This effect was due to the higher potential of the 1:1 substrate type, finer in texture, to runoff fine material.
The existence of superficial coarse fragments in mine reclamation plots, acting like a rock mulch, is related to decreases in erosion (Evans and Loch 1996), less susceptibility to interrill erosion, and stability from rill erosion (Sheridan et al. 2000b). However, the results of Poesen and Ingelmo-Sanchez (1992) demonstrate that a cover of rock fragments can cause increases in runoff and erosion when the coarse fragments are partially embedded in a surface crust. This was the effect observed in this experiment; the rock fragments, which were in approximate proportions of 45% in the 1:1 substrate type and 60% in the 1:2 substrate type, were incorporated in the substrate crust, and consequently, they were not effective in protecting against erosion.
The effect of substrate type on runoff N[O.sub.3.sup.-]-N concentration was significant in plots of Sets I and II, where 1:1 substrate type plots had higher N[O.sub.3.sup.-]-N runoff concentration than substrate type 1:2 plots. The same tendency was observed in Set III, but not in set IV were the high slope gradient probably disguised these differences. The 1:1 substrate type averaged 138 mg/kg whereas the 1:2 type averaged 74 mg/kg. The influence on EC was lower; the 1:1 substrate type averaged 4.2 dS/m 25[degrees]C, whereas the 1:2 type averaged 3.1 dS/m 25[degrees]C.
In general, nitrogen and sediment runoff were above permissible rates; therefore, some measures such as the design of terraces retaining runoff would be important in the reclamation of these slopes to avoid downstream water pollution.
Results of the laboratory germination experiment were: A. halimus 68%, B. scoparia 49%, G. maritima 14%, O. milliacea 91%, P. lentiscus 10%, R. sphaerocarpa 6%, S. kali 53%, and S. vermiculata 72%. From these results is possible to assume a low success in the field of some species (G. maritima, P. lentiscus, and R. sphaerocarpa).
In the field plots, vegetation began to emerge in April 1999. This is be considered quite late, probably due to the severely dry conditions of the year, and moreover because the crust formed after the first rains retarded emergence. The critical crust strength values that prevent emergence depend on crust and plant characteristics, but values >1 kg/[cm.sup.2] limit emergence of several species (Hillel 1972). The crust formed usually reached surface strength values >1 kg/[cm.sup.2]; plants could only germinate when soil moisture reserves were adequate, and emerge once the strength of the crust was decreased when it was moistened enough by the rain.
Statistical analysis of emergent plants per [m.sup.2] in the microplots (Fig. 9) indicated that Set III had highest emergence (not significantly different from Set II). There were no differences between Sets I and II, and Set IV had the lowest emergence. Substrate crusting was not responsible for these differences, since there were no surface strength differences between Sets III and IV; nor was superficial moisture content responsible, as this was not different for Sets II, III and IV. These observations also allow us to consider that the trapping of seeds in the crust was not responsible of these differences of emergence rates.
[FIGURE 9 OMITTED]
Consequently, the differences in emergence were not due to substrate properties and could be due to failures of seed fixation, which seemed to be related to the slope gradient. A limiting slope gradient of >46 to 56% (gradients of Sets III and IV) seemed prevent seed fixation. Wind and runoff can be considered as responsible agents for seed removal from the plots, and substrate crusting probably enhanced seed movement.
The effect of substrate type, sowing type, and position on emergence within sets of plots is exhibited in Table 7 (except for Set IV, in which emergence was <1 plant/[m.sup.2]). Substrate type differences were not significant. The effect of position was significant in plots of Sets I and III; lower positions had higher emergence than upper ones. These results could not be assessed based on difference of surface strength or superficial moisture properties from those in upper plot positions. It suggests movement of seeds downslope on the plots, probably due to runoff. But there were no differences in seed emergence between upper and lower positions in plots of Set II, due to its south aspect, making it more protected from the wind. This fact would indicate that wind could also influence seed removal.
The number of species emergening was very low; only 4 species were successful and these species were included in the sowing type R (and colonised the other microplots sown with different seed mixtures). None of the species belonging to the A sowing type emerged (the sowing type F had particular problems as explained below). Viability of purchased seeds included in the A-mixture (and in the hydrosowing mixture, F-mixture) is `guaranteed', so it is likely that substrate properties such as salinity or low moisture retention caused this failure. Moreover, some of the indigenous hand-collected species used in the A mixture showed a low emergence rate in the laboratory experiments.
The colonisation of microplots that were not sown with the R sowing type was due to both natural seed dispersion from the surroundings and seed movement from sown plots elsewhere. The highest emergence detected in the control (non-sowed) microplots in Set III was remarkable, suggesting that Set III had especially favourable topographic conditions to receive natural seed dispersion although it has higher slope than Sets I and II.
Percentage cover in microplots was already measurable at the beginning of April (day of year 102), and maximum cover was reached in August (day of year 220); after that vegetation dried and declined (Fig. 10). Sets I and III reached the highest covers (considering the average of all microplots R, A, B, T and F), higher than the others sets during the periods of the maximum cover (cover of plots in Set IV remained <10% and was always significantly lower than the other sets of plots a result of the very low emergence). Coverage of Sets I, II, and III can be considered acceptable, since the levels of vegetative cover achieved in many mine reclamation sites would range from 20 to 60% (Loch 2000a).
[FIGURE 10 OMITTED]
There were no differences in coverage between Sets I and III. This indicated that the best conditions for plant growth were reached in Set I, because it had the highest cover without the highest emergence. This was probably related to its lower slope, which allowed more water retention and higher substrate nitrogen content. These results are similar to those of Navas et al. (1999), who found that slope gradients up to 35% were negatively correlated with biomass amount in semi-arid badland soils treated with biosolids.
Substrate thickness presumably was an important condition enhancing plant growth allowing higher water retention and nutrient availability. Sets I and III, which were thicker, reached higher cover, although this effect was difficult to confirm because it could not be compared with Set IV, due to the low emergence, nor with Set II, where the south aspect caused drier conditions and thus reduced plant growth.
The effect of substrate type, sowing type, and position on total percentage cover within sets of plots is shown in Table 8. Vegetation cover was influenced by substrate type, the higher nutrient content in the substrate type 1:1, and probably its higher water retention appeared to be the cause of the higher vegetation cover.
The effect of position on total percentage cover was significant in plots of Set I, where lower positions had higher percentage cover than upper positions probably related to the emergence rate. In contrast cover in Set III was similar in the upper and lower microplots, but emergence was much higher in lower parts, which suggests a limit to plant development around these cover percentages.
R-type microplots were the most successful and F-type the least successful (it was noted that birds and ants ate these seeds and the seedlings dried out in a few days). As R was the most successful sowing type, its percentage cover over time is displayed in Fig. 11. The shape of the curves was very similar to those observed in Fig. 10 (average of R, A, B, T and F), but the cover reached values about 25% higher in all the sets except in Set IV.
[FIGURE 11 OMITTED]
The R-type was even better than the type T, which had the same seed mixture but double the amount of seeds. The difference between sowing type R and T was the sowing date; therefore, it confirmed the idea that in this climate winter sowings are more successful that spring ones.
Due to the monitoring dates, the effects of vegetation on runoff properties could not be checked, although some considerations can be made. Despite the scarce information on the effects of vegetative cover on runoff and erosion for reclaimed mined land, some authors suggest values of target cover levels that would provide protection against erosion under specific conditions.
Loch (2000b), in experiments with 30% slope, concludes that a cover to protect against erosion from overland flows is adequate if >60%, but inadequate if <35%. He also reported that plots with a mean cover >42-47% will not rill, and are therefore unlikely to produce high rates of erosion. According to Carroll et al. (2000), substrate erosion rates rapidly declined once the vegetation covered more than half of the plot in 10-30% slope experiments.
To conclude, Loch (2000a) considers, in 15% slope plots, that a level of 50% grass cover could be set as a target for stable rehabilitation, at least when stoloniferous grasses are a major component of the vegetation. He also explains that the potential impacts of cover type on erosion are likely to be important.
Looking at the percentage cover of the sowing type R, it is possible to indicate that the level of colonisation in Sets I, II, and III was probably effective to control erosion, at least from June (day of year 165) onwards in Set I, and from about July onwards in Sets II and III, the period when the cover was >50%. It is important to note that the slope in this experiment was higher, and the kind of vegetation was different (bush cover), to that the previous works (grass cover).
The kind of successful species were saline and nitrophilous semi-arid adapted bushes, and only 4 species had success: A. halimus, B. scoparia, Lolium spp., and S. kali. All these species were included in the R and T sowing types (except Lolium spp., which was not sown, it came mixed in the straw), and also appeared in the other microplots (A, B, F). No species of the type A or F mixture emerged in any microplot. In order to simplify these results, only the effect of sets of plots, substrate type, and position on the relative percentage cover of individual species was analysed. The comparison of relative average percentage cover of individual species among sets of plots (Fig. 12) showed that B. scoparia was the dominant species in Sets I and III, mainly due to marked nitrophilous character. S. kali and A. halimus (important because it is a perennial species) establishment was better in Set II, and these species are well adapted to salinity and lack of water (Naidu and Harwood 1997), remembering that Set II is the most saline due to its southerly aspect. A. halimus and Lolium spp. cover was <10% in all sets. Lolium spp. reached a higher cover in Set I than the other sets, probably due to its higher water retention that favours the development of grasses.
[FIGURE 12 OMITTED]
Fig. 13 shows the effect of substrate type on the relative average percentage cover of individual species. The substrate type 1:1 was colonised more by B. scoparia and Lolium spp. than the substrate type 1:2 because of its higher nitrogen availability. S. kali growth was higher in 1:2 substrate type plots, probably because its development decreases with high phosphorous levels and organic amendments (Johnson 1998). No substrate type effect was found in A. halimus development.
[FIGURE 13 OMITTED]
Fig. 14 shows the effect of position on the relative average percentage cover of individual species. There were no differences in the cover reached by A. halimus and Lolium spp. between the upper and lower positions. B. scoparia reached a higher cover in lower positions, whereas S. kali was more successful in upper positions, probably related again to substrate water availability.
[FIGURE 14 OMITTED]
Further studies on the evolution of substrate properties and vegetation characteristics are required to achieve a more definitive conclusion, since these properties may change with time.
The experiments developed to reclaim the slope mine spoil banks in the Mequinenza coalfied (Ebro basin) showed that the rehabilitation techniques employed could be successful, but there were some limitations.
Crusting was a problem in the plots, caused by the silty texture of the substrate. The rock fragments, expected to protect against erosion acting as mulch, were incorporated in the crust and consequently they were not effective.
Emergence was delayed by crusting, measures to control it might be employed to avoid this degradation. In the lower parts of the plots the traffic of machinery should be avoided.
The slope gradient and the slope aspect determined the superficial substrate moisture after a rain event during the emergence period. South set aspect and increasing slope gradient probably induced lower moisture content. When the slope was >33% the lower positions had higher moisture levels than the upper ones.
Sediment runoff was over permissible rates due to loose material from the fine spoil. Nitrogen runoff concentration was also excessive in the runoff. The design of terraces retaining runoff water is a key point in the reclamation of these slopes to avoid downstream water pollution.
Substrate EC and mainly N[O.sub.3.sup.-]-N dynamics could not be explained using the data recovered in this experiment, since factors such as plant absorption, and nitrification or denitrification, have an important influence, and they were not measured.
Emergence rate was in general acceptable (but limited to very adapted species), except in Set IV, where it was very low. The failure of emergence was due to problems of seed fixation more than problems of seedling emergence. The slope gradient was the main cause. Indeed, in wind expose slopes a limiting slope gradient of >46 to 56% seemed not to allow seed fixation. This could have some important implications since it supposes the necessity of using a rougher substrate to improve seed fixation. Topographic conditions, such as those which conditioned wind impact, played an important role in the revegetation of these spoil banks allowing easier seed arrival and fixation.
The differences in emergence depending on sowing date confirmed the idea that, in these climatic conditions, winter sowings are more successful than spring sowings. The sowing type comprising indigenous saline and nitrophilous semi-arid adapted species was the most successful. Probably (temporary data are not enough to confirm this), the plant colonisation in all the sets, except in the set with 56% slope, was effective to control erosion, at least from June or July on, which was the period with cover >50%.
The best substrate type was 1:1 type, but a proper terrace design retaining sediment and runoff is necessary to avoid downstream pollution.
The best conditions for plant establishment were reached in the plots with the lowest slope (30%), the shortest length (13 m), and with westerly aspect. Plots with slope gradient of 56% were impossible to colonise above 10%. These observations will help decisions about future dump shapes.
Figure 2. Diagram of the experimental plots. Each set had two randomised replications of substrate type 1 : 1 and substrate type 1 : 2. Set Aspect Slope Plot gradient (%) length (m) I West 30 13.0 II South 33 17.5 III West 46 17.5 IV West 56 13.5 Table 1. Fine earth characteristics of fine and coarse spoils pre pig slurry addition Characteristic Fine Coarse pH (1:2.5) 7.6 7.5 ECe (dS/m at 25[degrees]C) 6.32 6.27 Organic matter (%) 13.9 23.4 CaC[O.sub.3] equivalent (%) 51 52 Olsen P ([micro]g/g) 3 2 Potassium AcN[H.sub.4] (mg/kg) 42 33 Kjeldahl nitrogen (%) 0.21 0.28 Sand (A) (%) 16.6 62.3 Silt (A) (%) 63.3 27.6 Clay (A) (%) 20.1 10.1 (A) USDA particle size distribution (Soil Survey Division Staff 1993) Table 2. Effect of substrate type and position on surface strength (kg/[cm.sup.2]) within sets of plots Values are the average of 108 measures. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Set Substrate Position SxP (A) 1:1 1:2 Up Down I 6.3a 7.2a 6.0b 7.6a n.s. II 5.1b 7.0a 5.0b 7.1a n.s. III 8.8a 6.9b 7.4a 8.3a n.s. IV 7.9a 8.4a 5.9b 10.4a ** ** P < 0.01; n.s., not significant. (A) Substrate x position interaction. Table 3. Effect of substrate type and position on superficial moisture (%) 48 h after a rainfall event during emergence period within sets of plots Values are the average of 4 measures. Interactions were not significant. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Set Substrate Position 1:1 1:2 Up Down I 7.8a 6.5a 6.6a 7.7a II 5.3a 5.2a 4.8a 5.7a III 6.8a 5.6a 4.0b 8.4a IV 5.8a 4.7a 3.7b 7.1a Table 4. Effect of substrate type and position on substrate nitrate concentration (mg/kg dry-fine earth basis) within sets of plots Values are the average of 28 measures. Interactions were not significant. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Set Substrate Position 1:1 1:2 Up Down I 341a 177b 246b 288a II 278a 164b 245a 198b III 271a 147b 212a 150b IV 273a 144b 159b 235a Table 5. Effect of substrate type on substrate electrical conductivity (dS/m at 25[degrees]C 1:1 soil to water extract) within sets of plots Values are the average of 28 measures. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Substrate Position SxP Set 1:1 1:2 Up Down I 8.8a 7.1b 8.4a 7.5a n.s. II 8.5a 7.4b 7.4b 8.5a ** III 8.0a 7.3b 7.1b 8.2a n.s. IV 7.6a 7.2a 7.2a 7.6a n.s. ** P < 0.01; n.s., not significant. Table 6. Effect of substrate type on runoff sediment yield (g/L), nitrate concentration (mg/kg), and electrical conductivity (dS/m 25[degrees]C) within set of plots Values are the average of 7 measures. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Set Sediment yield N[O.sup-.sub.3]-N EC Substrate: 1:1 1:2 1:1 1:2 1:1 1:2 I 28.3a 15.0a 234a 92b 6.6a 3.2b II 31.6a 33.5a 122a 43b 3.3a 3.0a III 41.1a 29.9a 110a 65a 4.0a 2.7a IV 62.0a 23.6b 90a 100a 3.0a 3.6a Table 7. Effect of substrate type, sowing type and position on plants emergence (plants/[m.sup.2]) within sets of plots Values are the average of 12 measures for substrate type and sowing position and 8 measures for sowing type. Interactions were not statistically significant. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Set Substrate Sowing type Position 1:1 1:2 R A B Up Down I 6a 7a 12a 2b 3b 3b 8a II 8a 6a 13a 5b 4b 8a 7a III 11a 11a 12a 5a 16a 2b 20a IV Without statistical analysis due to the low emergence Table 8. Effect of substrate type, sowing types and position on total percentage cover within sets of plots Analysis of position is made considering only showing type R, A and B that had both lower and upper positions. Values are the average of 92 measures for substrate type; 48 measures for sowing types R, A, and B; 20 measures for sowing types T and F; and 72 measures for sowing position. Interactions were not significant. Within rows and factors, values followed by the same letter are not significantly different (Duncan's multiple range test at P = 0.05) Set Substrate Sowing Position 1:1 1:2 R A T F B Up Down I 33a 22a 50a 22b 45a 3c 26b 24b 41a II 26a 12b 34a 20b 7cd 2d 16bc 25a 21a III 31a 22b 38a 30ab 20b 0f 24ab 30a 32a IV Without statistical analysis due to the low percent cover
We thank the Spanish Education Ministry, and the Department of Environment (Catalonia Government), for their financial support. The authors wish to acknowledge the co-operation of Carbonifera del Ebro company, the assistance of J. Casas and the collaboration of S. Kampf.
Bernal MP, Roig A, Garcia D (1993) Nutrient balances in calcareous soils after application of different rates of pig slurry. Soil Use and Management 9, 9-14.
Besnier F (1989) `Semillas: biologia y tecnologia.' (Mundi-Prensa: Madrid)
Cameron KC, Rate AW, Carey PL, Smith NP (1995) Fate of nitrogen in pig effluent applied to a shallow stony pasture soil. New Zealaand Journal of Agricultural Research 38, 533-542.
Campbell DJ, O'Sullivan F (1991) The cone penetrometer in relation to trafficability, compaction, and tillage. In `Soil analysis. Physical methods'. (Eds KA Smith, CE Mullins) pp. 399-429. (Marcel Dekker, Inc. Publishing: New York)
Carroll L, Merton L, Burger P (2000) Impact of vegetative cover and slope on runoff, erosion and water quality for field plots on a range of soil and spoil materials on central Queensland coal mines. Australian Journal of Soil Research 38, 313-327.
Cerda A, Garcia-Fayos P (1997) The influence of slope angle on sediment, water and seed losses on badland landscapes. Geomorphology 18, 77-90.
Coyne MS, Zhai Q, Mackown CT, Barnishel RI (1998) Gross nitrogen transformation rates in soil at a surface coal mine site reclaimed for prime farmland use. Soil Biology and Biochemistry 30, 1099-1106.
Evans KG (2000) Methods for assessing mine site rehabilitation design for erosion impact. Australian Journal of Soil Research 38, 231-247.
Evans KG, Loch RJ (1996) Using the RUSLE to identify factors controlling erosion rates of mine soils. Land Degradation and Development 7, 267-277.
Felton GK (1995) Temporal variation of soil hydraulic properties on municipal solid waste amended mine soils. Transactions of the American Society of Agricultural Engineers 38, 775-782.
Folk L (1951) A comparison chart for visual percentage estimation. Journal of Sedimentary Petrology 21, 32-33.
Garcia-Fayos P, Recatala TM, Cerda A, Calvo A (1995) Seed population dynamics on badland slopes in southeastern Spain. Journal of Vegetation Science 6, 691-696.
Hillel D (1972) `Optimizing the soil physical environment toward greater crop yields'. (Academic press: New York)
Hornick SB (1988) Use of organic amendments to increase the productivity of sand and gravel spoils: effect on yield and composition of sweet corn. American Journal of Alternative Agriculture 3, 156-162.
Johnson EA, Fryer GL (1992) Physical characteristics of seed microsites-movement on the ground. Journal of Ecology 80, 23-36.
Johnson NC (1998) Responses of Salsola Kali and Panicum virgatum to mycorrhizal fungus, phosphorous and soil organic matter: implications for reclamation. Journal of Applied Ecology 35, 86-94.
Leavitt KJ, Fernandez CJ, Nowak RS (2000) Plant establishment on angle of repose mine waste dumps. Journal of Range Management 53, 442-452.
Leiros MC, Gil-Sotres F, Trasar-Cepeda MC, Saa A, Seoane S (1996) Soil recovery at the Meirama opencast lignite mine in northwest Spain: a comparison of the effectiveness of cattle slurry and inorganic fertilizer. Water, Air, and Soil Pollution 91, 109-124.
Loch RJ (1996) Using rill/interrill comparisons to infer likely responses of erosion to slope length. Implications for land management. Australian Journal of Soil Research 34, 489-502.
Loch RJ (2000a) Effects of vegetation cover on runoff and erosion under simulated rain and overland flow on an rehabilitated site on the Meandu Mine, Tarong, Queensland. Australian Journal of Soil Research 38, 299-312.
Loch RJ (2000b) Using rainfall simulation to guide planning and management of rehabilitated areas: Part I. Experimental methods and results from a study at the northparkes mine, Australia. Land Degradation and Development 11, 221-240.
Nacci S, Pla I (1992) Estudio de la resistencia a la penetracion de los suelos con equipos de penetromia desarrollados en el pais. Agronomia Tropical 42, 115-132.
Naidu BP, Harwood MR (1997) Opportunities for landscape stabilisation and revegetating disturbed lands in stressful environments with exotic or native forages. Tropical Grasslands 31, 364-369.
Navas A, Machin J, Navas B (1999) Use of biosolids to restore the natural vegetation cover on degraded soils in the badlands of Zaragoza (NE Spain). Bioresource Technology 69, 199-205.
Poesen J, Ingelmo-Sanchez F (1992) Runoff and sediment yield form topsoils with different porosity as affected by rock fragment cover and position. Catena 19, 451-474.
Porta J, Lopez-Acevedo M, Rodriguez R (1986) `Tecnicas y experimentos en edafologia.' (Col-legi Oficial d'Enginyers Agronoms de Catalunya: Barcelona)
SAS Institute (1989) `SAS/STAT User's Guide Version 6.' 4th edn (SAS Institute: Cary, NC)
Schroeder SA (1987) Slope gradient effect on erosion of reshaped spoil. Soil Science Society of America Journal 51, 405-409.
Sheridan GJ, So HB, Loch RJ, Pocknee C, Walker CM (2000a) Use of laboratory-scale rill and interrill erodibility measurements for the prediction of hillslope-scale erosion on rehabilitated coal mine soils and overburdens. Australian Journal of Soil Research 38, 285-297.
Sheridan GJ, So HB, Loch RJ, Walker CM (2000b) Estimation of erosion model erodibility parameters from media properties. Australian Journal of Soil Research 38, 265-284.
Smith SJ, Schepers JS, Porter LK (1990) Assessing and managing agricultural nitrogen losses to the environment. Advances in Soil Science 14, 1-33.
Soil Survey Division Staff (1993). `Soil survey manual.' US Department of Agriculture, Natural Resources Conservation Service Handbook 18. (US Government Printing Office: Washington DC)
Sopper WP (1992) Reclamation of mine land using municipal sludge. Advances in Soil Science 17, 351-431.
Vieillefon J (1979) Contribution a l'amdioration de l'etude des sols gypseux. Cahiers Orstom, Serie Pedologie XVII, 195-223.
Wetselaar R, Smith GD, Angus JF (1998) Field measurement of soil nitrate concentrations. Communications in Soil Science and Plant Analysis 29, 729-739.
Wilcox BP, and Wood MK (1989) Factors influencing interrill erosion from semiarid slopes in New Mexico. Journal of Range Management 42, 66-70.
Ye ZH, Wong JWC, Wong MH, Lan CY, Baker AJM (1999) Lime and pig manure as ameliorants for revegetating lead/zinc mine tailings: a greenhouse study. Bioresource Technology 69, 35-43.
Manuscript received 21 November 2000, accepted 4 January 2002
M. Salazar, R. M. Poch, and A. D. Bosch
Departament de Medi Ambient i Ciencies del Sol, Escola Tecnica Superior d'Enginyeria Agraria, Universitat de Lleida, Alcalde Rovira Roure, 177. Lleida 25198, Spain; e-mail: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Salazar, M.; Poch, R.M.; Bosch, A.D.|
|Publication:||Australian Journal of Soil Research|
|Date:||Sep 1, 2002|
|Previous Article:||Measurement and prediction of permethrin persistence in six Malaysian agricultural soils.|
|Next Article:||A pedo-transfer function (PTF) for estimating soil bulk density from basic soil data and its comparison with existing PTFs.|
|Impact of vegetative cover and slope on runoff, erosion, and water quality for field plots on a range of soil and spoil materials on central...|
|TUNISIA - The Geology.|